基于希尔伯特-黄变换的径流规律分析

希尔伯特－黄变换在处理非线性、非平稳、弱相依、高度复杂的时间序列时，比传统的时间序列分析方法更具优势。通过对时间序列进行经验模态分解（EMD），得到若干反映时间序列演化过程的本征模态函数（IMF）和序列变化的残余项（趋势项），然后对每个本征模态函数作希尔伯特－黄变换，得到反映径流规律的统计特征值，从而获得序列变化的周期。对渭河临潼水文站的年径流观测数据进行分析，得出年径流变化存在9．52年、13．70年、21．74年和41．67年的周期，且有逐渐减小的趋势。

Runoff Law Analysis by Hilbert-Huang Transform

Compared with traditional analytical method of time series, Hilbert - Huang Transform has advantages to describe time - frequency characteristics for nonlinear, unsteady, weak - dependency and highly complex time series. Empirical Mode Decomposition （EMD） method is adopted to analyze time series, Intrinsic Mode Function （IMF） is reflected and the evolution of time series and sequence changes in the residual items （trend term） is obtained. Then, Hilbert - Huang transform is made for each IMF. The statistics reflected in runoff and sequence changes cycle are obtained. Taking hydrological stations of Lintong in the Wei River as examples, periods analysis of IMF shows that runoff change has period of 9.52,13.70,21.74 and 41.67 years, and gradually decreases.